The identification of appropriate reference genes used for normalization and whose expression is not affected by the experimental factors is a crucial step in gene expression studies performed using real-time PCR (a.k.a. quantitative PCR, or RT-qPCR). The choice of unsuitable reference genes can compromise the results and may lead to wrong interpretations of biological phenomena.
The literature is constellated by studies focusing on the selection of reference genes in different plant species, tissues, conditions [1
]; a list of potential candidates whose orthologs can be tested in the plant/tissue of choice is usually provided. Several types of software are available to assess the stability of a set of candidate reference genes [5
] and a ranking is provided to show those most stable. The comparison of expression patterns in plants subjected to a combination of abiotic stresses [9
], or tissues showing strong differences in terms of composition and developmental stage, poses a challenge when data have to be normalized. In these circumstances, if a tissue maximization design is used [10
], then reference genes suitable for all the conditions/tissues studied need to be identified. This might require the screening of several candidates, given the heterogeneity of the experimental conditions tested.
One example of experimental heterogeneity is the study of gene expression in the stems of fibre crops, like textile hemp (Cannabis sativa
L.), since they are characterized by tissues differing dramatically in their composition [11
]. The stems of hemp are composed of cortical tissues harbouring cellulose-rich sclerenchyma fibres (which mechanically support the phloem, a.k.a. bast fibres) and a woody core (often referred to as hurd or shiv). These tissues can be easily isolated, since the cortex can be peeled off and subsequently processed to separate the bast fibres from the epidermis/parenchyma/collenchyma. While the core fibres are lignified and characterized by the typical secondary cell wall layers S1-S2-S3, bast fibres possess a gelatinous layer (G-layer; Figure 1
) composed of crystalline cellulose which is similar to that found in tension wood [12
]. The G-layer of bast fibres, however, does not exert the same contractile function as in tension wood [13
]. It should be noted that hemp stems, unlike the other fibre crop flax (Linum usitatissimum
), possess also secondary bast fibres which are shorter and more lignified than primary fibres and originate from the cambium [11
Besides the differences in tissue composition, hemp stems additionally show a basipetal lignification gradient: younger internodes at the top are indeed rapidly elongating and the bast fibres show a relatively thin cell wall, while older internodes at the base of the stem cease elongation, synthesize a thick secondary cell wall and are more lignified. The transition from elongation to cell wall thickening is marked by an empirically-determined point, called the “snap point” [14
The heterogeneous lignification of its stem tissues makes hemp very interesting as a model to study cell wall-related processes: the inner/outer tissues of the same internode, as well as those collected from younger/older internodes along the stem, can be studied to carry out high-throughput gene expression profiling. Such studies allow addressing molecular questions on, e.g., the regulation of bast fibre cell wall formation and maturation.
Some studies are available on the identification of reference genes in other fibre crops, notably flax (Linum usitatissimum
], kenaf (Hibiscus cannabinus
] and jute (Corchorus capsularis
]. These studies on one hand show that care should be taken when choosing the reference genes, as their stability might change depending on the tissue analysed and the treatment studied, and on the other hand they highlight that the different algorithms available for the evaluation of reference genes may not generate an identical ranking list (see Section 2.2
No specific study focused on reference genes for RT-qPCR data normalization in hemp is yet available to our knowledge. Given the importance of hemp as a multi-purpose crop for different industrial applications, it is desirable to identify candidate reference genes to perform gene expression analyses. In this work we fill this gap by providing a list of candidates suitable for RT-qPCR data normalization in hemp stem tissues (i.e., bast fibres and shivs). Additionally, we validate them by performing an expression profiling of genes involved in the provision of precursors for lignin biosynthesis. More specifically, we chose genes coding for transaldolase isoforms of the pentose phosphate pathway, TRA1 and TRA2, which synthesize erythrose 4-phosphate, and 3-deoxy-d-arabino-heptulosonate 7-phosphate synthase (DHS) 1 and 2, diverting C-skeletons towards the shikimate pathway.
Given the importance of hemp woody and bast fibres in the textile and biocomposite sectors [18
], our study is a useful guide for future molecular studies centred on this economically important fibre crop.
In this study, 12 candidate reference genes were analysed to test their stability and we validate their use in data normalization on hemp stems. The data shown highlight that it may be necessary to select different reference genes when heterogeneous biological samples are studied, as is the case of the contrasting hemp stem tissues. The studied reference genes can not only be used on textile hemp varieties, but also represent candidates to test on oil and drug Cannabis
varieties. Additionally, the 12 reference genes here reported can eventually be tested in expression studies focused on close plant species, as for example the Cannabaceae member H. lupulus
, given the overall sequence homology (Figure S3
). The expression analysis of genes involved in the non-oxidative phase of the pentose phosphate pathway and the shikimate pathway, notably TRA1
have also been studied. It was possible to identify isoforms potentially involved in lignification. Further investigations are required to define the roles of other genes playing in the different cell wall-related processes of the hemp stem: lignification requires a complex network of metabolic pathways and signals, different to those required for bast fibre synthesis. It will be interesting in the future to study them functionally. For example, our study opens up the way to future studies centred on the pentose phosphate pathway, which is a central metabolic pathway for both the primary and secondary plant metabolism. The role of glucose-6-phosphate dehydrogenase, a central player in the pathway [32
], can be addressed to understand its contribution to the shunt of precursors needed for lignin biosynthesis.